Projects

MSU Award Number

361478-191001-021000

Current NGI File Number

21-NGI4-11

Associated MSU Award or NGI File Numbers

20-NGI-111 (361458-191001-021000)

Start Date

10/1/2021

End Date

2024-9-30 0:0:0

Project Status

Completed

Award Amount

$137,508.00

Project Title

Video and Imagery Data Management for Fisheries Assessments

PI Last Name

Mickle

PI First Name

Paul

PI Affiliation

MSU

Co-PI Last Name

Prior

Co-PI First Name

Jack

Co-PI Affiliation

MSU

Research Theme

Data Management, Ecosystem Management

Funding Source

NMFS

Project Description

Artificial Intelligence (AI) is being employed to help make digital media, that were collected for fisheries assessments, to be more assessable and usable. Researchers are organizing the media, which are housed at various NOAA NMFS laboratories and partnering agencies in the southeast US, and using AI to automate the detection and identification of fish species from video. They are compiling and archiving image libraries and associated manual annotations, collecting and digitizing historic video, and comparing human-based annotations with those generated by Video Image Analytics for the Marine Environment (VIAME) algorithms, which speeds up video processing time ad quality control procedures. The annotation library and model iterations for fish detection and identification are available for viewing and use on Kitware’s VIAME web application. The researchers share results with NOAA fisheries labs to demonstrate how to use the fish tracking models and the VIAME applications for their project needs. The model’s ability to classify and track reef fish has improved, especially for high priority species such as snappers and groupers.